MVCC means "Multiversion Concurrency Control" and yes, that's pretty much exactly how it works (creating copies of the rows, so that different snapshots see different rows). In practice it's more complicated of course.…
I don't think XML was all that widely used, and the implementation looked more "looks interesting, let's develop something and people will use that". With JSONB it's more like "people are already using JSON data type…
I think that if you're using JSONB for data that can be easily stored in a proper tabular form (aka relation), you're doing it wrong. No doubt there will be people using it this way, though, and maybe there are cases…
People are working on other storage types for PostgreSQL. That's all I can say at this moment.
Yeah, IRC is a good place to ask too. The obvious problem is that not all developers hang there all the time, so for longer discussions the mailing lists are better (and you can just attach test cases, for example).
PostgreSQL does not have in-place update in the first place, on every UPDATE it does a DELETE+INSERT. So "partial update" does not make much sense, because you'll create a new copy of the row anyway.
I really wonder what you mean by UPSERTS. Because the traditional meaning for UPSERT is "UPDATE if exists, INSERT otherwise", aka MERGE. And you're right PostgreSQL is missing that, but hopefully we'll get something…
It doesn't have MySQL-like pluggable storage, and frankly I'm thankful for that. How many of the MySQL storage engines actually work, including transactions for queries working with tables using different storage…
Expression indexes are nice, and you can still do that. It however requires you to know what you'll be looking for (and create index on that particular expression). The stuff I find really awesome is indexing for…
Why? What mistake? I mean, many people won't use that, but for many it's very handy. And the features delivered with JSONB (e.g. the indexing) is way better than what was available with XML.
JSONB had issue with compression, because the compression is decided on the fly by checking if first part of the serialized data, and in the original format the first part was incompressible (thus no compression was…
The only response I have to that is "talk to developers on the mailing list" (either pgsql-performance or pgsql-hackers). Maybe it's possible to improve the planning - maybe your queries are uncommon / difficult to…
IANAL, of course, but ... 1. There's no suitable legal entity that could receive the copyright/rights on behalf of the PostgreSQL project. There are entities handling domains/... but that's not really useful. 2. The…
I don't really see how that matters. This only shows how utterly broken the concept of "patch Tuesdays" is ... If you plan your internal deployment updates based on the belief that the schedule will never change, then…
Well, they thought so too. And it turned not to be true. And now they're blaming the others.
If this is used just as an AP, how much harm can a baseband backdoor do? I mean, there are no interesting sensors, the other side already knows the (approximate) location ... assuming the user uses proper crypto (not on…
If you start making exceptions, everyone will want one.
You're just citing the press release (or a blog post, or whatever). How does that respond to the original question? The issue mentioned in the article [https://code.google.com/p/google-security-research/issues/de...]…
If you define a limit for disclosure, and then not stick to it, why to define a limit in the first place? 90 days is more than enough - if MS has a lot of internal overhead, you should probably complain to them, not to…
So, who forces Microsoft to stick to (so called) patch Tuesdays? No one, actually - it's Microsoft internal schedule, and clearly there are cases when it's absolutely unreasonable - e.g. when there's a 0day in the wild.…
That really depends on what version you're using now, and what exactly you mean by scalability. If you're using 9.1 or older, you may see a significant improvement in OLTP workloads on many-core machines (making it…
I'm not a big fan of MongoDB, but I don't think the introduction of JSONB in PostgreSQL 9.4 makes it a dud (which does not mean MongoDB is not a dud for other reasons). JSONB allows you to do a lot of things that people…
We certainly are. We're operating an analytical service operated on PostgreSQL - tens of TBs of data, hundreds of machines, tens of thousands of clients. Initially it was running on MySQL, but because of various reasons…
AFAIK the default serializer does not guarantee the order to be stable, but I haven't participated in the development of this feature that closely. So maybe I'm wrong. Anyway, it should be possible to write a simple…
Yeah, it's old. Most basic principles described by Bruce are still valid, though. But if you're looking for fresh info on PostgreSQL tuning, buy this: https://www.packtpub.com/big-data-and-business-intelligence/... It's…
MVCC means "Multiversion Concurrency Control" and yes, that's pretty much exactly how it works (creating copies of the rows, so that different snapshots see different rows). In practice it's more complicated of course.…
I don't think XML was all that widely used, and the implementation looked more "looks interesting, let's develop something and people will use that". With JSONB it's more like "people are already using JSON data type…
I think that if you're using JSONB for data that can be easily stored in a proper tabular form (aka relation), you're doing it wrong. No doubt there will be people using it this way, though, and maybe there are cases…
People are working on other storage types for PostgreSQL. That's all I can say at this moment.
Yeah, IRC is a good place to ask too. The obvious problem is that not all developers hang there all the time, so for longer discussions the mailing lists are better (and you can just attach test cases, for example).
PostgreSQL does not have in-place update in the first place, on every UPDATE it does a DELETE+INSERT. So "partial update" does not make much sense, because you'll create a new copy of the row anyway.
I really wonder what you mean by UPSERTS. Because the traditional meaning for UPSERT is "UPDATE if exists, INSERT otherwise", aka MERGE. And you're right PostgreSQL is missing that, but hopefully we'll get something…
It doesn't have MySQL-like pluggable storage, and frankly I'm thankful for that. How many of the MySQL storage engines actually work, including transactions for queries working with tables using different storage…
Expression indexes are nice, and you can still do that. It however requires you to know what you'll be looking for (and create index on that particular expression). The stuff I find really awesome is indexing for…
Why? What mistake? I mean, many people won't use that, but for many it's very handy. And the features delivered with JSONB (e.g. the indexing) is way better than what was available with XML.
JSONB had issue with compression, because the compression is decided on the fly by checking if first part of the serialized data, and in the original format the first part was incompressible (thus no compression was…
The only response I have to that is "talk to developers on the mailing list" (either pgsql-performance or pgsql-hackers). Maybe it's possible to improve the planning - maybe your queries are uncommon / difficult to…
IANAL, of course, but ... 1. There's no suitable legal entity that could receive the copyright/rights on behalf of the PostgreSQL project. There are entities handling domains/... but that's not really useful. 2. The…
I don't really see how that matters. This only shows how utterly broken the concept of "patch Tuesdays" is ... If you plan your internal deployment updates based on the belief that the schedule will never change, then…
Well, they thought so too. And it turned not to be true. And now they're blaming the others.
If this is used just as an AP, how much harm can a baseband backdoor do? I mean, there are no interesting sensors, the other side already knows the (approximate) location ... assuming the user uses proper crypto (not on…
If you start making exceptions, everyone will want one.
You're just citing the press release (or a blog post, or whatever). How does that respond to the original question? The issue mentioned in the article [https://code.google.com/p/google-security-research/issues/de...]…
If you define a limit for disclosure, and then not stick to it, why to define a limit in the first place? 90 days is more than enough - if MS has a lot of internal overhead, you should probably complain to them, not to…
So, who forces Microsoft to stick to (so called) patch Tuesdays? No one, actually - it's Microsoft internal schedule, and clearly there are cases when it's absolutely unreasonable - e.g. when there's a 0day in the wild.…
That really depends on what version you're using now, and what exactly you mean by scalability. If you're using 9.1 or older, you may see a significant improvement in OLTP workloads on many-core machines (making it…
I'm not a big fan of MongoDB, but I don't think the introduction of JSONB in PostgreSQL 9.4 makes it a dud (which does not mean MongoDB is not a dud for other reasons). JSONB allows you to do a lot of things that people…
We certainly are. We're operating an analytical service operated on PostgreSQL - tens of TBs of data, hundreds of machines, tens of thousands of clients. Initially it was running on MySQL, but because of various reasons…
AFAIK the default serializer does not guarantee the order to be stable, but I haven't participated in the development of this feature that closely. So maybe I'm wrong. Anyway, it should be possible to write a simple…
Yeah, it's old. Most basic principles described by Bruce are still valid, though. But if you're looking for fresh info on PostgreSQL tuning, buy this: https://www.packtpub.com/big-data-and-business-intelligence/... It's…